U.S. patent application number 13/261294 was filed with the patent office on 2013-01-17 for wafer polishing method.
The applicant listed for this patent is Shunsuke Mikuriya, Shinichi Ogata, Kazushige Takaishi, Keiichi Takanashi, Tetsurou Taniguchi. Invention is credited to Shunsuke Mikuriya, Shinichi Ogata, Kazushige Takaishi, Keiichi Takanashi, Tetsurou Taniguchi.
Application Number | 20130017763 13/261294 |
Document ID | / |
Family ID | 44115078 |
Filed Date | 2013-01-17 |
United States Patent
Application |
20130017763 |
Kind Code |
A1 |
Takaishi; Kazushige ; et
al. |
January 17, 2013 |
WAFER POLISHING METHOD
Abstract
An object of the present invention is to provide a method of
polishing silicon wafers, capable of suppressing generation of
undesired sounds from carriers and reducing the thickness variation
of the wafers after polished. The method is a wafer polishing
method in which wafers 20 are polished by supplying a polishing
solution to surfaces 30a of a pair of polishing pads 30 positioned
above and below carriers 10 each having a circular hole 11 for
retaining the wafers 20, the carriers 10 being thinner than the
wafers 20; and sliding the polishing pads 30 relatively to the
carriers 10, thereby simultaneously polishing both surfaces of the
wafers 20 retained in the carriers 10. The method is characterized
in that information sourced from the carriers 10 when a difference
between the thickness of the carriers 10 and the thickness of the
wafers 20 reaches a predetermined value is detected to calculate
the thickness of the wafers 20, thereby terminating polishing.
Inventors: |
Takaishi; Kazushige; (Tokyo,
JP) ; Takanashi; Keiichi; (Tokyo, JP) ;
Taniguchi; Tetsurou; (Tokyo, JP) ; Ogata;
Shinichi; (Tokyo, JP) ; Mikuriya; Shunsuke;
(Tokyo, JP) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
Takaishi; Kazushige
Takanashi; Keiichi
Taniguchi; Tetsurou
Ogata; Shinichi
Mikuriya; Shunsuke |
Tokyo
Tokyo
Tokyo
Tokyo
Tokyo |
|
JP
JP
JP
JP
JP |
|
|
Family ID: |
44115078 |
Appl. No.: |
13/261294 |
Filed: |
November 30, 2010 |
PCT Filed: |
November 30, 2010 |
PCT NO: |
PCT/JP2010/071777 |
371 Date: |
May 23, 2012 |
Current U.S.
Class: |
451/5 |
Current CPC
Class: |
B24B 37/08 20130101;
H01L 21/02024 20130101; B24B 37/042 20130101; B24B 37/013
20130101 |
Class at
Publication: |
451/5 |
International
Class: |
B24B 37/013 20120101
B24B037/013 |
Foreign Application Data
Date |
Code |
Application Number |
Dec 1, 2009 |
JP |
2009-273255 |
Dec 4, 2009 |
JP |
2009-276517 |
Claims
1. A wafer polishing method comprising the steps of: supplying a
polishing solution to a surface of a pair of polishing pads
positioned above and below a carrier which has a circular hole for
retaining a wafer and is thinner than the wafer; and sliding the
polishing pads relatively to the carrier, thereby simultaneously
polishing both surfaces of the wafer retained in the carrier,
wherein information sourced from the carrier when a difference
between a thickness of the carrier and a thickness of the wafer
reaches a predetermined value, is detected to calculate the
thickness of the wafer and thereby terminate polishing.
2. The wafer polishing method according to claim 1, wherein the
termination of polishing is performed at a termination point of
polishing determined by extracting a first signal and a second
signal that are signals of two setup frequency ranges from the
information obtained from the carrier, and detecting the
termination point of polishing where an intensity of the first
signal is higher than a predetermined value and an intensity of the
second signal is lower than a predetermined value.
3. The wafer polishing method according to claim 1, wherein the
polishing solution is a polishing solution free of abrasive
grains.
4. The wafer polishing method according to claim 1, wherein the
information sourced from the carrier is sounds and/or
vibrations.
5. The wafer polishing method according to claim 1, wherein the
difference between the thickness of the carrier and the thickness
of the wafer at a time when the information is sourced, is in the
range of 0.1 pm to 20 pm.
6. The wafer polishing method according to claim 1, wherein a
pressure applied to the polishing pads in the direction
perpendicular to the surfaces of the wafer is in the range of 100
g/cm.sup.2 to 300 g/cm.sup.2.
7. The wafer polishing method according to claim 1, wherein the
information sourced from the carrier is detected with detectors
provided for each carrier.
8. The wafer polishing method according to claim 7, wherein the
detector is provided on a mechanism for holding the carrier.
9. The wafer polishing method according to claim 8, wherein the
detector is provided on external gears for moving the carrier.
10. The wafer polishing method according to claim 4, wherein the
detection of the sounds from the carrier is performed by detecting
sounds having a frequency of 10 Hz to 1000 Hz.
11. The wafer polishing method according to claim 1, wherein the
polishing solution contains a specific high molecular weight
component.
12. The wafer polishing method according to claim 2, wherein the
extracting the signals from the information is calculation of a
power spectra of digital data obtained by A/D converting the
detected information and of a mean value or a maximum value of
signal intensities of power spectra corresponding to the setup
frequency ranges.
Description
TECHNICAL FIELD
[0001] The present invention relates to a wafer polishing method
and, in particular to a method of simultaneously polishing both
surfaces of wafers by rotating carriers retaining the wafers
between an upper polishing plate and a lower polishing plate
provided with a polishing pad on each surface.
RELATED ART
[0002] For a process of polishing silicon wafers, a polishing
method of forcing an upper polishing plate and a lower polishing
plate that are provided with a polishing pad on their surface
against wafers retained by carriers and moving the wafers between
the upper polishing plate and the lower polishing plate that are
rotated, thereby simultaneously polishing both surfaces of the
wafers has been generally employed. The carriers each have a
circular hole for retaining the wafers, and have a smaller
thickness than the wafers. In this polishing method, a polishing
solution containing abrasive grains has been generally used. This
polishing method has been used because surfaces and rear surfaces
of a plurality of wafers can be simultaneously polished and
accordingly productivity and flatness of the surfaces and the rear
surfaces of the wafers can be improved.
[0003] The above polishing method using carriers and a polishing
solution containing abrasive grains however has a problem in that
the carriers are worn by the abrasive grains contained in the
polishing solution, and the thicknesses of the carriers vary, which
would lead to thickness variation of the wafers after polished.
[0004] Patent Document 1 discloses carriers for a double-side
polishing apparatus capable of preventing the carriers from wearing
at the time of simultaneously polishing both surfaces of wafers.
However, the carriers for a double-side polishing apparatus in
[0005] Patent Document 1 are carriers each made by coating the
surface of a resin matrix with a DLC (diamond-like carbon) film.
Therefore, the production cost is extremely high as compared with
metal carriers, carriers each having a metal plate of which surface
and rear surface are coated with resin, and resin carriers, which
are commonly used. This has been a problem.
[0006] In cases where a polishing solution free of abrasive grains
is used as a polishing solution for simultaneously polishing both
surfaces of wafers, whichever of the following carriers may be used
to prevent the carriers from wearing: metal carriers, carriers each
having a metal plate of which surface and rear surface are coated
with resin, and resin carriers. However, as polishing proceeds, the
difference between the thickness of the carriers and the thickness
of the wafers becomes small, and polishing pads contact the
carriers to cause the carrier to vibrate, which results in
undesired sounds generated from the carriers. This has been another
problem.
[0007] When polishing is performed in a state where carriers
continue to vibrate, not only the problem of undesired sounds being
generated from the carriers arises, but the variation of the wafer
thickness that should be reduced by polishing would conversely
increase as a result of the variation in the contact of polishing
pads with both surfaces of the wafers, which is yet another
problem. Therefore, the time point at which the carriers begin to
vibrate should be ensured to be detected, and polishing is required
to be terminated at the time point.
[0008] Patent Document 2 discloses a method of simultaneously
polishing both surfaces of wafers by moving the wafers retained by
carriers between upper and lower polishing plates that are rotated.
In the polishing method, from vibrations generated from the
polishing plates accompanying wafer polishing, one or a plurality
of specific frequencies at which the vibration level changes
reflecting the progression of polishing are selected in advance,
and the change in the vibration level at the specific frequencies
is detected in polishing, so that the termination point of
polishing is estimated from the change in the vibration level.
[0009] Further, in other polishing methods in which polishing is
terminated with the use of polishing sounds generated when wafers
are polished, polishing sounds generated in polishing of wafers are
detected, and the termination point of polishing is determined
based on a change in the polishing sounds, for example as disclosed
in Patent Documents 3 and 4. In accordance with this method, the
termination point of polishing can be determined during the
progress of a polishing process using a relatively simple system
configuration.
[0010] Patent Documents
[0011] Patent Document 1: Japanese Patent Application Publication
No. 2006-303136
[0012] Patent Document 2: Japanese Patent Application Publication
No. 2005-252000
[0013] Patent Document 3: Japanese Patent Application Publication
No. H6-45299
[0014] Patent Document 4: Japanese Patent Application Publication
No. 2001-15467
DISCLOSURE OF THE INVENTION
Problems to be Solved by the Invention
[0015] However, in the invention of Patent Document 2, in fact, a
vibration sensor provided on an upper surface of an upper polishing
plate detects not only vibrations of polishing plates but also the
total vibrations including vibrations of the polishing plates,
vibrations of wafers, and vibrations of carriers. Therefore, noise
is mixed in the detected specific frequencies at which the
vibration level changes; as a result, the change in the vibration
level cannot be accurately measured. Thus, the termination point of
polishing cannot be accurately detected, which would lead to
further variations in the thickness of the wafers polished.
[0016] Further, in the inventions of Patent Documents 3 and 4,
while the termination point can be quantitatively determined, cases
of polishing using carriers are not considered. Thus, the thickness
variation of wafers after polished due to wear of the carriers
still remains as a problem to be solved.
[0017] An object of the present invention is to provide a method of
polishing silicon wafers, capable of suppressing undesired sounds
generated from carriers and reducing the thickness variation of the
wafers after polished.
Means for Solving the Problem
[0018] In order to solve the above problems, the inventors of the
present invention have made various studies on a wafer polishing
method including the steps of supplying a polishing solution to a
surface of a pair of polishing pads positioned above and below a
carrier which has a circular hole for retaining a wafer and is
thinner than the wafer; and sliding the polishing pads relatively
to the carrier, thereby simultaneously polishing both surfaces of
the wafer retained in the carrier. As a result, the inventors found
the following facts. The carrier generates information such as
vibrations or sounds when the difference between the thickness of
the carrier and the thickness of the wafer reaches a specific
value. And by detecting the information, the thickness of the wafer
can be determined. Furthermore, the termination of polishing based
on the determined thickness allows the polishing amount of the
wafer to be stably controlled; therefore, undesired sounds
generated from the carrier can be effectively suppressed, and a
wafer after polished with reduced thickness variation can be
obtained.
[0019] In order to achieve the above object, the present invention
primarily includes the following components.
[0020] (1) A wafer polishing method comprising the steps of:
supplying a polishing solution to a surface of a pair of polishing
pads positioned above and below a carrier which has a circular hole
for retaining a wafer and is thinner than the wafer; and sliding
the polishing pads relatively to the carrier, thereby
simultaneously polishing both surfaces of the wafer retained in the
carrier. Information sourced from the carrier when a difference
between a thickness of the carrier and a thickness of the wafer
reaches a predetermined value, is detected to calculate the
thickness of the wafer and thereby terminate polishing.
[0021] (2) The wafer polishing method according to (1) above,
wherein the termination of polishing is performed at a termination
point of polishing determined by extracting a first signal and a
second signal that are signals of two setup frequency ranges from
the information obtained from the carrier, and detecting the
termination point of polishing where an intensity of the first
signal is higher than a predetermined value and an intensity of the
second signal is lower than a predetermined value.
[0022] (3) The wafer polishing method according to (1) above,
wherein the polishing solution is a polishing solution free of
abrasive grains.
[0023] (4) The wafer polishing method according to (1) above,
wherein the information sourced from the carrier is sounds and/or
vibrations.
[0024] (5) The wafer polishing method according to (1) above,
wherein the difference between the thickness of the carrier and the
thickness of the wafer at a time when the information is sourced,
is in the range of 0.1 pm to 20 .mu.m.
[0025] (6) The wafer polishing method according to (1) above,
wherein a pressure applied to the polishing pads in the direction
perpendicular to the surfaces of the wafer is in the range of 100
g/cm.sup.2 to 300 g/cm.sup.2.
[0026] (7) The wafer polishing method according to (1) above,
wherein the information sourced from the carrier is detected with
detectors provided for each carrier.
[0027] (8) The wafer polishing method according to (7) above,
wherein the detector is provided on a mechanism for holding the
carrier.
[0028] (9) The wafer polishing method according to (8) above,
wherein the detector is provided on external gears for moving the
carrier.
[0029] (10) The wafer polishing method according to (4) above,
wherein the detection of the sounds from the carrier is performed
by detecting sounds having a frequency of 10 Hz to 1000 Hz.
[0030] (11) The wafer polishing method according to (1) above,
wherein the polishing solution contains a specific high molecular
weight component.
[0031] (12) The wafer polishing method according to (2) above,
wherein the extracting the signals from the information is
calculation of a power spectra of digital data obtained by A/D
converting the detected information and of a mean value or a
maximum value of signal intensities of power spectra corresponding
to the setup frequency ranges.
Effect of the Invention
[0032] The present invention can provide a polishing method capable
of suppressing undesired sounds generated from carriers and further
reducing thickness variation of wafers after polished.
BRIEF DESCRIPTION OF THE DRAWINGS
[0033] FIG. 1 is a diagram for illustrating a state of polishing
wafers.
[0034] FIG. 2 is a cross-sectional view for illustrating a state of
members in wafer polishing with respect to one embodiment of a
wafer polishing method in accordance with the present
invention.
[0035] FIGS. 3(a) and 3(b) are graphs showing the intensity of
information generated from carriers. Specifically, FIG. 3(a) is a
graph showing the intensity of sounds generated from the carriers,
and FIG. 3(b) is a graph showing the intensity of vibrations
generated from the carriers.
[0036] FIG. 4 is a graph showing the relationship between the
elapsed time (second) and the intensity (A.U.) of signals (first
signal and second signal) of two frequency ranges, extracted from
the information sourced from carriers in polishing.
[0037] FIG. 5 is a graph showing the relationship between the
elapsed time (second) and the intensity (m/s.sup.2) of signals of
vibrations generated from carriers in polishing.
[0038] FIG. 6 is a graph showing the relationship between polishing
time (min) and thickness variation (%) of carriers with respect to
a case of wafer polishing using a polishing solution free of
abrasive grains and a case of wafer polishing using a polishing
solution containing abrasive grains.
BEST MODE FOR CARRYING OUT THE INVENTION
[0039] A wafer polishing method of the present invention will be
described with reference to the drawings.
[0040] A wafer polishing method of the present invention is a
method in which wafers 20 are polished by supplying a polishing
solution (not shown) to surfaces 30a of a pair of polishing pads 30
positioned above and below carriers 10 each having a circular hole
11 for retaining the wafers 20, the carriers 10 being thinner than
the wafers 20; and sliding the polishing pads 30 relatively to the
carriers 10, thereby simultaneously polishing both surfaces of the
wafers 20 retained in the carriers 10 as shown in FIG. 1.
[0041] The method is characterized in that information sourced from
the carriers 10 when a difference between the thickness of the
carriers 10 and the thickness of the wafers 20 reaches a
predetermined value, is detected to calculate the thickness of the
wafers 20, and thereby terminate polishing.
[0042] With the above structure, as shown in FIG. 2, information
sourced from the carriers 10 when a difference between the
thickness TC of the carriers 10 and the thickness TW of the wafers
20 reaches a predetermined value (in FIG. 2, information sourced
due to vibrations of the carriers 10 in the directions indicated by
the arrows) is detected to calculate the thickness of the wafers
20, thereby making it possible to appropriately terminate
polishing. Thus, control of the polishing amount of the wafers 20
can be ensured, so that the wafers 20 having been polished with
reduced thickness variation can be obtained.
[0043] FIGS. 3(a) and 3(b) are graphs showing the intensity of
information (herein referring to sounds and vibrations) for both
cases where carriers are and are not generating information (sounds
and vibrations). FIG. 3(a) shows data of sounds generated from the
carriers. FIG. 3(b) shows data of vibrations generated from the
carriers. The portions enclosed by broken lines in FIGS. 3(a) and
3(b) show the situations where sounds and vibrations are generated
from the carriers, respectively. The relevant portions show that
the intensities of sounds and vibrations are higher than those in
cases where the information is not sourced. Therefore, for example,
the sounds and vibrations generated when the thickness of the
wafers 20 is desirable are perceived in advance, and polishing is
terminated when the relevant sounds and vibrations are achieved.
Thus, polishing can be accurately controlled.
[0044] Further, the termination of polishing is preferably
performed at a termination point of polishing determined by
acquiring the information sourced from the carriers 10 when the
difference between the thickness of the carriers 10 and the
thickness of the wafers 20 reaches a predetermined value, and
extracting a first signal and a second signal that are signals of
two setup frequency ranges from the information obtained, to detect
the termination point of polishing where the intensity of the first
signal is higher than a predetermined value and the intensity of
the second signal is lower than a predetermined value.
[0045] The first signal and the second signal extracted from the
information refer to two signals of different frequency ranges that
are extracted, after acquisition of the information sourced from
the carriers 10, from the acquired information by a specific
process. In the present invention, a signal of high frequency is
referred to as a first signal, while a signal of low frequency is
referred to as a second signal. Note that the frequencies of the
first signal and the second signal are not particularly limited and
may be any value as long as the termination of polishing can be
controlled by a combination of the first signal and the second
signal.
[0046] FIG. 4 is a graph showing the relationship between the
elapsed time (second) and the intensity (A.U.) of the signals
(first signal and second signal) of the two frequency ranges,
extracted from information (vibrations) sourced from the carriers
10 in polishing. As is also clear from FIG. 4, until undesired
sounds are generated from the carriers, the intensities of both the
first signal and the second signal fluctuate; however, while the
intensity of the first signal is high at the point where undesired
sounds are generated, the intensity of the second signal remains
low. Therefore, setting the intensity (predetermined value) which
the first signal should exceed and the intensity (predetermined
value) which the second signal should fall behind allows the
termination point of polishing to be quantitatively determined. On
the other hand, FIG. 5 is a graph showing the relationship between
the elapsed time (second) and the intensity (m/s.sup.2) of the
signals of vibrations generated from the carriers in polishing.
When only vibration signals are detected, many peaks of similar
intensity are shown (0 to 480 seconds, 700 seconds to 1000
seconds), which makes it impossible to ascertain the timing when
undesired sounds are generated from the above- mentioned carriers.
Thus, the termination point of polishing cannot be determined.
[0047] The information sourced from the carriers 10 is not
particularly limited and may be, for example, sounds, vibrations,
or heat generated from the carriers as long as the thickness TW of
the wafers can be determined based on the information. However, the
information is preferably sounds and/or vibrations generated from
the carriers 10 because the termination point of polishing the
wafers 20 can be easily and reliably detected.
[0048] Note that the method of extracting signals from the
information is not particularly limited as long as the
above-mentioned first signal and second signal can be effectively
extracted with the method. For example, since desired signal data
can be obtained relatively easily, it is preferable to perform the
signal extraction by calculating power spectra of digital data
obtained by A/D converting the detected information, and by
calculating a mean value or a maximum value of signal intensities
of power spectra corresponding to the setup frequency ranges. And
if necessary, only specified frequencies may be A/D converted by
filtering the information, or the peak hold processing may be
conducted after such filtering.
[0049] Further, since the polishing amount can be more accurately
controlled, a difference between the thickness TC of the carriers
10 and the thickness TW of the wafers 20 (TC-TW) at the time when
information is sourced from the carriers 10, is preferably in the
range of 0.1 .mu.m to 20 .mu.m although depending on the rotation
speed of the polishing pads 30, pressure, and the type of the
polishing solution as well.
[0050] The information (sounds or vibrations and the like) sourced
from the carriers 10 can be detected, for example with detectors 40
provided adjacent to the carriers 10 as shown in FIG. 1. Here, the
detectors 40 may be, for example, a sound input microphone 41 and
vibration detectors 42 shown in FIG. 1.
[0051] When the information to be detected is sounds, their
frequencies are preferably in the range of 10 Hz to 1000 Hz. The
frequency range is expected to be suitable for finding the
termination point of polishing; therefore, if polishing is
terminated with sounds outside the frequency range, desired wafer
thickness would not be achieved.
[0052] Further, the frequency range of the first signal is more
preferably in the range of 200 Hz to 1000 Hz, while the frequency
range of the second signal is more preferably in the range of 10 Hz
to 200 Hz. The first signal and the second signal of these
wavelength ranges make it possible to ensure determination of the
termination point of polishing when undesired sounds are generated
from the carriers 10 because their signal intensities greatly
differ from one another.
[0053] Meanwhile, when vibrations generated from the carriers 10
are detected, as shown in FIG. 1, the vibration detectors 42 are
preferably provided on a mechanism (in FIG. 1, external gears 50,
which are members used for moving the carriers 10) for holding the
carriers 10 to detect the information sourced from the carriers 10.
This is because it is difficult to detect the vibrations of the
carriers 10 itself being interposed between the polishing pads 30,
and it is easier to detect the vibrations of the external gears 50,
thereby ensuring easy acquisition of the information from the
carriers 10.
[0054] Further, the detected sounds and/or vibrations generated
from the carriers 10 are converted into signals, and the signals
are sent to a control device 60 as shown in FIG. 1, so that the
first signal and the second signal are extracted using the control
device 60; whether to continue or terminate polishing can also be
determined in this manner.
[0055] The carriers 10 used in a polishing method of the present
invention are, as shown in FIG. 1, members for retaining the wafers
20 to be polished, which each include a circular hole for retaining
the wafers 20. The thickness TC is smaller than the thickness TW of
the wafers 20. The shape and properties of the carriers 10 are not
particularly limited, and carriers commonly used for polishing both
surfaces of wafers may be used. The carriers 10 preferably have
such properties as high wear resistance, a small coefficient of
friction with polishing cloth, and high acid resistance or high
alkali resistance. The carriers 10 may be made of, for example, a
fiber reinforced plastic, that is, a combination of resin such as
epoxy resin, phenol resin, or polyimide resin and reinforcement
fiber such as glass fiber, carbon fiber, or aramid fiber.
[0056] The polishing pads 30 are members for sliding the wafers 20
being interposed between the upper and lower polishing pads to
simultaneously polish the surfaces of the wafers 20, as shown in
FIG. 1 and FIG. 2. The polishing pads 30 preferably have properties
such as a small coefficient of friction with the wafers 20 and the
carriers 10. For example, an urethane material having a hardness
(Shore D) of 75 to 85 with a compressibility of 2% to 3% may be
used. When the polishing solution does not contain abrasive grains
therewithin, the polishing pads 30 are preferably polishing pads
having fixed abrasive grains. Fixed abrasive grains here refer to
solid bodies made of a different material from the polishing pad
matrix that are fixed to a surface of the polishing pads, and the
abrasive grains used may be made of, for example, a ceramic such as
silica or alumina; diamond, silicon carbide, or a compound thereof;
or a high molecular weight polymer such as polyethylene or
polypropylene. Further, their shapes are not limited to particulate
form, and may be in the form of solid, gel, or the like.
Furthermore, kinds of silica that can be used may be, for example,
prepared by either a dry process (combustion process/arc process)
or a wet process (sedimentation process/sol-gel process).
[0057] A polishing solution used in a polishing method of the
present invention is preferably a polishing solution free of
abrasive grains. Use of a polishing solution containing abrasive
grains makes it difficult to detect information from the carriers
10 that is required for the termination of polishing. Here, the
"polishing solution free of abrasive grains" is polishing solution
in which abrasive grain components may be contained but such
abrasive grain components are not positively added.
[0058] FIG. 6 is a graph showing the relationship between polishing
time (min) and thickness variation (%) of carriers in cases of
polishing wafers using a polishing solution containing abrasive
grains and a polishing solution free of abrasive grains. FIG. 6
shows that the carriers are more worn out in the case of using the
polishing solution containing abrasive grains.
[0059] The type of the polishing solution is not particularly
limited. For example, an alkaline aqueous solution with pH
controlled to the range of 8 to 14 is preferably used. With a pH of
less than 8, the etching action is weak to reduce the polishing
rate; on the other hand, with a pH of more than 14, handling of the
polishing solution becomes difficult. The alkaline aqueous solution
may be an aqueous ammonia solution, or an alkaline hydroxide
solution such as potassium hydroxide or sodium hydroxide, or an
alkaline carbonate solution. Alternatively, an aqueous solution of
hydrazine or amines can be used. It is preferable to use an aqueous
solution of amines in particular in terms of increasing the
polishing rate. Further, the polishing solution preferably contains
a specific high molecular weight component, since undesired sounds
(squeals) generated from the carriers 10 can be further reduced.
The type of the high molecular weight component is not particularly
limited, but hydroxyethyl cellulose or polyethylene glycol is
preferably used. In particular, hydroxyethyl cellulose, which has a
relatively high molecular weight, can efficiently reduce the
coefficient of friction between the polishing pads 30 and the
carriers 10 due to the bearing action between the polishing pads 30
and the carriers 10. As a result, defects in wafers after polished
can be reduced. Further, the concentration of the high molecular
weight component added to the polishing solution is preferably in
the range of 0.01 ppm to 1000 ppm. If the concentration of the high
molecular weight component is less than 0.01 ppm, the friction
coefficient in polishing is high, which would cause defects in the
surface of the wafers after polished. On the other hand, if the
concentration is more than 1000 ppm, the polishing rate is
significantly reduced, and considerable time is required for mirror
polishing.
[0060] In terms of removing metal ions contained in the polishing
solution, a chelating agent is preferably added to the polishing
solution. The chelating agent may be any substance without
particular limitation as long as it has a chelating effect on metal
ions. For example, a phosphonic acid type chelating agent, an
aminocarboxylic acid type chelating agent, or the like can be used.
Considering the solubility in an alkaline aqueous solution, an
aminocarboxylic acid type chelating agent is preferable. In
addition, considering chelating effect on heavy metal ions, an
aminocarboxylic acid salt such as ethylenediaminetetraacetic acid
(EDTA) or diethylenetriaminepentaacetic acid (DTPA) is more
preferable. The concentration of the chelating agent is preferably
in the range of 0.1 ppm to 1000 ppm, which allows metal ions such
as Cu, Zn, Fe, Cr, Ni, and Al to be trapped.
[0061] Further, in a polishing method of the present invention, the
polishing pads 30 are slid relatively to the carriers 10, thereby
polishing surfaces of the silicon wafers 20. The sliding method is
not necessarily limited in particular. The sliding may be performed
by moving only the polishing pads 30 or by moving only the carriers
10; alternatively, the sliding may be performed by relatively
moving both the polishing pads 30 and the carriers 10.
[0062] The pressure applied to the polishing pads 30 in the
direction perpendicular to the surfaces of the wafers 20 is
preferably in the range of 100 g/cm.sup.2 to 300 g/cm.sup.2. When
the pressure is less than 100 g/cm.sup.2, the pressure is
excessively low, so that the information sourced from the carriers
10, such as sounds and vibrations would not be sufficiently
obtained. On the other hand, when the pressure is more than 300
g/cm.sup.2, the excessively high pressure makes sounds and
vibrations of the carriers 10 be easily generated; thus, the
polishing termination point of the wafers 20 would not be
accurately ascertained.
[0063] Note that the foregoing descriptions merely illustrate some
examples of embodiments of the present invention and various
modifications can be made thereto within the scope of claims.
EXAMPLES
Example 1
[0064] In Example 1, as shown in FIG. 1, silicon wafers 20 having a
diameter of 300 mm were polished by a wafer polishing method
including the steps of supplying a polishing solution free of
abrasive grains (a potassium hydroxide aqueous solution containing
ethylenediaminetetraacetic acid at a concentration of 10 ppm as a
chelating agent, and hydroxyethyl cellulose at a concentration of 1
ppm as a high molecular weight component) to a surface of a pair of
polishing pads 30 made of urethane positioned above and below
carriers 10 each having a circular hole 11 for retaining the wafers
20, the carriers 10 being 750 pm in thickness, which were thinner
than the wafers 20; and sliding the polishing pads relatively to
the carriers, thereby simultaneously polishing both surfaces of the
wafers retained in the carriers. The termination point of polishing
was determined by acquiring the vibrations generated from the
carriers 10 using vibration detectors 42 provided on the external
gears 50; extracting a first signal (a mean value of 200 Hz to 1000
Hz) and a second signal (a mean value of 10 Hz to 100 Hz) from the
obtained vibrations; and detecting the timing when the intensity of
the first signal is equal to or more than a predetermined value
(0.02 A.U.) and the intensity of the second signal is equal to or
less than a predetermined value (0.02 A.U).
[0065] The polishing conditions (the type of the polishing
solution, whether the polishing solution contains a high molecular
weight component or not, and the detection method of a polishing
termination point) are shown in Table 1. Further, the silicon
wafers were polished five times in total (25 sheets).
Example 2
[0066] In Example 2, silicon wafers (25 sheets) were polished by a
similar method to Example 1 except for that the sounds generated
from the carriers 10 were acquired by a sound input microphone 41
provided adjacent to the carriers 10, and a first signal (480 Hz)
and a second signal (100 Hz) were extracted from the sounds
obtained.
[0067] The polishing conditions (the type of the polishing
solution, whether the polishing solution contains a high molecular
weight component or not, and the detection method of a polishing
termination point) are shown in Table 1.
Comparative Example 1
[0068] In Comparative Example 1, silicon wafers (25 sheets) were
polished under similar conditions to Example 1 except for using a
polishing solution containing abrasive grains (a potassium
hydroxide aqueous solution containing colloidal silica as free
abrasive grains).
[0069] The polishing conditions (the type of the polishing
solution, whether the polishing solution contains a high molecular
weight component or not, and the detection method of a polishing
termination point) are shown in Table 1.
Comparative Example 2
[0070] In Comparative Example 2, silicon wafers (25 sheets) were
polished under similar conditions to Example 1 except for that
polishing was controlled not using the information (sounds or
vibrations) sourced from the carriers but using a setup time.
[0071] The polishing conditions (the type of the polishing
solution, whether the polishing solution contains a high molecular
weight component or not, and the detection method of a polishing
termination point) are shown in Table 1.
Evaluation Method
(1) Wear of Carriers
[0072] The thickness TC of the carriers in Example 1 and
Comparative Example 1 was measured before polishing, after a
predetermined time lapse from the beginning of the polishing, and
after the polishing. The wear level of the carriers was evaluated
by plotting the relationship between the polishing time and the
thickness TC of the carriers. The results are shown in FIG. 6 and
Table 1.
(2) Thickness variation of Wafers
[0073] The thickness profiles of the silicon wafers obtained in
Examples and Comparative Examples were measured using WaferSight
manufactured by KLA-Tencor Corporation, and the thickness variation
of the wafers was calculated.
[0074] Thus, the mean value of the variation was evaluated. The
evaluation results are as shown in Table 1, in which smaller values
are better.
TABLE-US-00001 TABLE 1 Detecting Evaluation method Thickness of
Polishing variation Polishing solution termination Wear of of Wafer
Type Polymer point Carrier (.mu.m) Example 1 Without Hydroxy-
Vibration Hardly 0.3 Abrasives ethyl from Carrier worn cellulose
(FIG. 6) Example 2 Without Hydroxy- Sounds -- 0.3 Abrasives ethyl
from Carrier cellulose Comparative With None Vibration Badly 0.5
Example 1 Abrasives from Carrier worn (FIG. 6) Comparative Without
Hydroxy- Time -- 0.6 Example 2 Abrasives ethyl cellulose
[0075] Table 1 and FIG. 6 show that when polishing was performed by
polishing methods of Example 1 and Example 2, the carriers were
hardly worn as compared to polishing methods of Comparative Example
1 and Comparative Example 2, and thickness variation of the
obtained silicon wafers was small. On the other hand, it is shown
that the carriers were badly worn and the thickness variation was
large in the polishing method of Comparative Example 1 in which
polishing was performed using a polishing solution containing
abrasive grains and in the polishing method of Comparative Example
2 in which polishing was controlled using a set time.
INDUSTRIAL APPLICABILITY
[0076] The present invention can provide a polishing method capable
of suppressing undesired sounds generated from carriers and further
reducing thickness variation of silicon wafers after polished.
Reference Numerals
[0077] 10: Carrier
[0078] 11: Circular Hole
[0079] 20: Wafer/Silicon Wafer
[0080] 30: Polishing Pad
[0081] 40: Detector
[0082] 41: Sound Input Microphone
[0083] 42: Vibration Detector
[0084] 50: External Gear
[0085] 60: Control Device
* * * * *